Professional Documents
Culture Documents
AutoInobile
Suspensions
COLIN CAMPBELL
M.Sc., C.Eng., M.I.Mech.Engrs
LONDON
CHAPMAN AND HALL
First published 1981
by Chapman and Hall Ltd
11 New Fetter Lane, London EC4P 4EE
© 1981 C. Campbell
Softcover reprint of the hardcover 1st edition 1981
Campbell, Colin
Preface page ix
1 Wheels and tyres 1
2 Springs 22
3 Suspension principles 41
4 Suspension geometry 54
5 Conventional systems 77
6 Road -holding 105
7 Dampers 120
8 Pneumatic suspensions 129
9 Hydropneumatic suspensions 143
10 Interconnected and no-roll suspensions 165
11 A small FWD saloon car: Ford Fiesta S 185
12 A high-performance sports car: Porsche 928 198
Index 211
We have had enough of action and of motion we,
Roll'd to starboard, roll'd to larboard,
When the surge was seething free.
Alfred Lord Tennyson
Preface
1.1 Suspension
Ventilation
Wheel stud
Nominal or
mting
Pitch diameter
circle
diameter
Protuberance ----;~b1
13.
Tyrrell Fl McRae FSOOO Lotus Fl Sur t ees F2
13" • 16"reor 15" • 17 '" reor 10" front 13 " ,10" front
Wt 4·10 kg Wt 5·44 kg wt 3·34 kg WI 3-40 kg
Fig. 1.4 Centre lock wire wheels with octagonal hub caps.
'Elastic belt' 01
rubberized canvas
reinforcement. The average life of the early car tyre was about
2000 miles.
In the 1920s the cross-ply construction was developed in
which the tyre casing was built up by placing several plies of
parallel cords inside the mould, the cords passing at a bias
angle from the inside beaded edge to the outer beaded edge.
Alternate plies, however, had the bias angle (relative to the
circumferential centre line) alternate between a positive and a
negative angle. The all-important difference from the earlier
canvas reinforcement is that each separate sheet of ply has
parallel cords which are coated completely with a rubber
compound. Direct contact between adjacent cords is,
therefore, avoided. The alternating angles between the bias
plies and the circumferential centre line is usually between
20° and 30°. Large angles are chosen when comfort is the
primary consideration. Racing tyres are made with an angle of
about 20°, giving a harsh ride, but increased resistance to the
centrifugal forces associated with racing speeds.
Until 1948 when Michelin introduced the 'X' series radial
tyres, many tyre manufacturers believed there was little left
for their research-and-development departments to do but to
make minor refinements in construction techniques and to
70 series
60 series
SO series
1.4.2 Skidding
In the 1920s when tread patterns of the type shown in Fig.
1.11 were in vogue, it was known that ribbed tyres gave long
life while block patterns gave good resistance to skidding.
Until makers like Dunlop developed their Cornering Force
'"
V = 46 m.p.h. v = 80 m.p.h.
4· GROOVE RIB TREAD
Fig. 1.13 Aquaplaning: water depth, 0.5 mm (photographed from below through transparent road surface).
16 Wheels and tyres
tested on good dry surfaces. To achieve maximum footprint
area, dry racing tyres are completely bald. They are, of
course, extremely dangerous on a wet road, since they
behave like surfboards. At speeds above about 50 m.p.h. the
buildup of water in front of the footprint acts like a wedge to
force a film of water across the whole area of the footprint,
'thus destroying the grip on the road. Ihis phenomenon is
known as aquaplaning (hydraplaning, in the USA). All good
modem tread patterns are designed to allow large volumes of
water to pass from front to rear of the footprint. The
influence of good footprint drainage is clearly illustrated in
Fig. 1.13.
Contact
3·0 500
2·8 400
2·6
2·4
300
~ 2·2
=
z
Q)
2·0
1·8
()
J...
.2 1· 6 200 kg
01
c
·c 1·4
Q)
cJ... 1·2
0
u '·0
0·8
0·6
0·4
0·2
Rim width. There is an optimum wheel rim width for any tyre
section. If a change is made to tyres with a lower aspect ratio,
it is advisable to change the wheels, too, since the use of
wide tyres on narrow rims introduces the danger of
destructive stresses in the side walls.
l f8 iI J
L I
J
Cornering
777~T~~~//~ force
Positive camber
,
I~ i l~
L i ,
Cornering
7777J»1!f17i77777777777777777? ~ force
Negative camber
CONTACT
PATCH
CORNERING ONLY
CORNERING +TRACTION
CORNERING + BRAKING
Reference
22 Tolerance to vibration
M,ofmoo_- p
X 1-----
--~~-----+--~Y4-~-+-L----~--------~~~
T = ~ = 21tJ~ . (2.1)
. . . - 21l'J
words as:
Mass of body (2.2)
Penodlc tIme - S· ffn f . .
tI ess 0 sprmg
Alternatively, if we take the deflection of the spring under
the weight of the body as d (the static deflection):
T = 21l'Ji (2.3)
where g, in this case, is the acceleration due to gravity.
21l')
a =d (T
2
(2.4)
r-~~ ~~.=~___.~
__ ._._J_ =t~Jl_m__m
__ : ______
Fig. 2.3
a = 0.05 x (-21t) 2
0.4
= 12.4 m/s 2
= 1.26 gravity.
--
.
'"
.........
\ \ \
'-... . _ . - '- - -- -.-..... "- ' - ' -
"-
\ ""'"
\
\
./ )
k=~ (~r
or, since T = 1// (where / is the frequency in Hz):
M
k= "4 (27Cf)2 N/m. (2.5)
Table 2.1 Spring rates and static deflections under 300 kg sprung mass
Spring rate k Static deflection d
Frequency f
(Hz) (N/mm) Obf/in) (mm) (in)
t""1"~---- S - - - - - - 1__-',
Fig. 2.5 Loss of contact in a ripple.
30 Springs
as much as 40 mm. Therefore, we can predict that a ripple
depth of about 150 mm is the limit for such a spring system,
if wheel contact is to be maintained.
If we soften the spring suspension to a frequency of
1.25 Hz, the static spring deflection is increased to 159 mm.
With this new spring the tyre will remain in full contact when
traversing the length of the ripple, even though the load on
the footprint will be reduced at the bottom of the ripple.
We have shown that our pogo stick system can be designed
to maintain reasonable surface contact with a ripple depth of
150 rom (6 in) with a relatively low frequency. There must,
however, be a lower limit to the length of ripple that can be
traversed at a given speed. In this analysis it is assumed that
the car is travelling at such a high speed that the inertia of the
relatively massive sprung body prevents any appreciable
vertical movement. The conditions leading to resonance
bouncing of the body will be discussed later. The natural
periodicity of the spring under the action of the unsprung
mass is not the same as that under the action of the sprung
mass. From Equation (2.1), we see that T ex: ,jM. Thus, with a
natural period of T under mass M f , t = T,j(m/M f ) when the
spring is oscillating under the smaller mass mf'
If we take the limiting ripple length as s metres and the car
velocity as v metres per second, the limiting ripple length is
given by:
Therefore
(2.6)
A close study of Tables 2.1 and 2.2 shows that a low spring
frequency helps the wheels to follow the contours of road
ripples at speeds of 100 k.p.h., but only if the ripples are
spaced at a greater distance than 9-14 m.
On the other hand, the use of a high spring frequency,
such as 2.5 Hz, will reduce these limits to 3.5-5.5 m,
depending upon the unsprung-to-sprung mass ratio, but the
limiting depth of the ripple will be correspondingly reduce~.
The choice of spring frequency is, therefore, not easy. Once
again, we are faced with that unattractive word 'compromise'.
Many modem designers are turning to rising rate suspensions
in an attempt to conquer this problem. Such spdngs give low
frequencies under small wheel movements and higher
frequencies under large movements.
s = lJ:.
Therefore
v =SfJ~.
If s is taken as 5 m:
v = 5 x 1.4J10
= 22 m/s
= 80 k.p.h. (49 m.p.h.).
A car with a smaller mfM ratio, say 1/12, and with the same
mean spring frequency, would travel most comfortably at a
speed of about 87 k. p.h. (54 m. p.h.).
Vintage sports cars with mean spring frequencies of about
2.0 Hz or even higher would be at a great disadvantage in
Movement of the sprung mass 33
such circumstances, since the optimum speed for comfort
would be much higher. Moreover, softly sprung modem cars
would have dug troughs in the road to a depth of at least
80 mm. The vintage sports car would be bumping on its
suspension stops, if the driver attempted to get 'into the
groove'. Naturally if only cars with stiff suspensions ever
used this road, the problem would not arise.
~8r-----~------'-----7r-----'
°O~~--~/~·O~--~2~·O~--~3~~~----47~
Frwquenc!I Ratio. ~I.fn.
Fig.2.6 Damping characteristics: note resonance whenfJfn is unity.
34 Springs
under the action of the sprung mass. Fig. 2.6 shows what
happens on a typical undamped suspension system when this
occurs. The undamped curve is indicated by the number 1.
Naturally on a washboard road the driver would accelerate
through the critical speed that caused body resonance. The
profile of a typical road surface, however, presents a wide
gamut of forCing frequencies and at any chosen cruising speed
a car with undamped springing would quite frequently strike
stretches of road that induced resonance.
Optimum damping is usually a compromise, a compromise
that is often influenced by the tester's individual preference.
Competition vehicles are sometimes fitted with adjustable
dampers and are set up to meet the requirements of the
particular driver. The contribution of the damper to
suspension design is of primary importance and will be
discussed fully in Chapter 7
1200
r::
S
..0
T
~ 800
z
It
fii
400
0+----.-----,----.----,----,----,----,
o 100 200 300 400 500 600 700
LOAD-Ibf
Fig. 2.9 Variation of stiffness with load and pressure for 5.60 x 13
cross-ply tyre.
1200
1000
"
~BOO
,i ///
T
~ 600
....
z
... /
~ 400
200",,//
1'0
0·1
k 3 =840
r-- ----...,
I m=l·n I
I c=305 I 0-0 Experimental.
- - - - Theoretical.
I ~100 k= 10 400 I
IL _ TY':!..T0de~ _ _
I
..J Pressure 15 Ib/in 2 •
Preload 500 Ibf.
System Excitation 27·9 Ibf.
0·01-j---.----,--.,--..,..-,----,--r_~_r___r---,r-_.__-T-_r__.
o 8 16 24 32 40 48 56
FREQUENCY -Hz
Fig. 2.11 Motion transmissibility across 5.60 x 13 cross-ply tyre.
1·0
0'1
~
0--0 Experimental
*]=840 - - - - Theoretical.
,---------,
I m-m I Pressure 15 Ib/in2 •
I tC-~15 I Preload 500 Ibf.
Excitation 27'9 Ibf.
I ~IOO *2=14200 I
I I
L _:!t.r!.!"~e~ ___ J
System
0'01 I I I I I I
o 8 10 24 ~2 40 48 50
FREQUENCY -Hz
Fig. 2.12 Motion transmissibility across 163 x 13 radial tyre.
References
Cen troid
K =Radius of"gyration
(3.3)
y--:!"""---
x
z
--~------~----------~B~-----Y
Fig. 3.5
y ~--~~--~------------+----------------y
E
'-1--- 5 -----1
Fig. 3.6
Suspension theory 47
Y--~-+~~~~~--~~J--------------------~Y
Fig. 3.7
b=_k_ 1 _
kl + k2 .
K2 is a quantity that can be estimated by the student or can
be measured accurately on special apparatus available to
manufacturers or at certain research establishments such as
the Cranfield Institute of Technology.
G can also be estimated in the drawing office (using such
computer facilities as may be available) or, if the vehicle
exists, by direct measurement of the axle weights on a
weigh bridge .
From the positions of C and G we know the value of x.
Substituting for p and q in Equation (3.5), Equations (3.7) and
(3.8) give us:
(r + x)(s - x) = c2 (3.9)
and substituting for s from Equation (3.6), we have:
(r + X)(~2 - x) = c 2
Hence:
Suspension theory 49
This gives two values for r. The negative root is, of course,
imaginary. There are also two values of s, a positive and a
negative. s can be found from Equation (3.6) by substituting
the value of r. It will be found that the negative value of s
equals the positive value of r and vice versa. The two double
conjugate points Hand J are now known.
(3.12)
(3.13)
Example 51
Periodic time in pitch
T
p
=
2
xJ(Restoring couple
Moment of inertia about G \
a bout for unit angular deflection}
G
(3.14)
(3.15)
3.4 Example
a = 1.224 m
b = 1.062 m
11 = a - X = 0.46 X 2.286 = 1.052 m
12 = b + X = 0.54 X 2.286 = 1.234 m
K2 = 1.05 X 1.052 X 1.234 = 1.363
X = a-II or 12 + b = 0.172 m
c2 = a X b = 1.3.
From this data we obtain the following:
Double conjugate point distance r = 1.269 m
Double conjugate point distance s = 1.074 m
Example 53
Front bounce frequency F f = 1.45 Hz
Rear bounce frequency F r = 1.69 Hz
Pitching frequency F p = 1.56 Hz.
This shows a general softening of the suspension.
For a small car a rear bounce frequency of about 1. 7 Hz is
very satisfactory. The only criticism we can make is that the
rear end frequency in bounce is very close to that in pitch.
The student could attempt a few experimental designs to see
if he can improve on the design of the Ford Motor Co. He
must always remember to work inside the physical limitations
imposed by the overall body design. For example, the use of
lower spring rates increases the total wheel travel. The
intrusion of the upper spring mountings into the engine bay
at the front, and the rear seating at the rear, could create
acrimony between the suspension engineer and the body
engineer.
Reference
4.1.1 Toe-in
This is a very small angle made by each front wheel plane
and the longitudinal axis of the car. In Fig. 4.1 the angle is
much exaggerated for the sake of clarity. The amount of
toe-in is usually measured as a difference in the distance
between right and left wheel rims at front and rear, both
measurements being made at hub level. A typical toe-in
would be as little as 3 mm. Without toe-in the inevitable
compliance (or 'free play' as a typical mechanic would call it)
in the several ball joints used in the steering linkages, could
easily lead to the phenomenon known as 'shimmy', where
the wheels flutter in and out within the limits set by the .~otal
compliance. The action of toe-in is simply to keep all the
steering linkages under tension. Toe-out would also take up
the slack and give stable straight-line running. The track rod
and other links would be under compression in this case.
This is undesirable with relatively long tubular links. If the
manufacturer's specified toe-in is exceeded, excessive tyre
wear will occur. The optional toe-in is usually established
experimentally.
Comber
Thearetica I center
of rotation of cambered
wheel
I --- ---- ---
\
I
\ \ I
\
\ \
\
!s:&
B
Understeered
c
Neutro I
A
Oversteered
Fig. 4.2 Effect of slip angles on directional stability. Vehicles A, B
and C are subjected to identical side forces. The right-hand car A
has greater slip angles at the rear than the front. This creates a
centrifugal force C acting in the same direction (apprOximately) as
the side force F. This calls for rapid steering correction by the driver.
The left-hand car C has greater slip angles at the front. In this case
the centrifugal force opposes the initial side force. This is a stable
situation. Neutral steer, as shown by car B, is given when slip
angles at front and rear are identical. In this case the car suffers a
sideways displacement, but does not steer to right or left.
Front wheel orientation 57
cornering power at the front relative to the rear. A car with
greater slip angles at the front than at the rear under the
action of a transient side force will be an understeering car. A
small degree of understeer is essential for good straight-line
stability. The reverse of understeer is called crversteer (see Fig.
4.2). This is discussed in more detail in Chapter 6.
Camber angle can change with suspension movement and
this is a critical design parameter when we come to layout
the geometry of the suspension linkages. Camber angle is
usually specified by the manufacturer in the unladen static
condition. Static angles seldom exceed two degrees.
Fig. 4.3 Details of kingpin and stub axle on beam front axle.
9°
Cas~or ° -----.I SwiveL-
angle --\ 7r--
\ -
- - - I~g lel
,0 --; '---,
Camber I :
I
angle I '~ il
· :
,I ,
i. (
11 3tl;i~ii6r--
Fig. 4.4 Modern steering layout, using upper and lower swivelling joints to replace kingpin.
Fro_nt wheel orientation 5(}
King pin
inclination
Theoretical plane
of wheel rotation
---
limit'.
6- _
_ Sliding
- - Racing
driver ~
5
'"~ ,I)o~,~
4 \
3
I
Skilled
driver
2
jt
/
/'.1'0
.,,/ oI'~
Average
driver
-100 -80 -60 -40 -20 60 80 100
Self- aligning torque (N m)
Fig. 4.6 Relationship between cornering force and self-aligning
torque on a typical tyre.
Front wheel orientation 61
Steering arm
Track rod
Ei3 ,
- I
------ _---:4--- ,
- I
,£:+3
Fig_ 4.7 Ackermann steering layout_
)¥-r%
, \
/ - , ,
. .-/ . ,.......... ---------- I
,/ "..___- I
/ I,
'/ I
. ../
r;i\
Roll
angle
~._F T,
G~I'
h '
" Mo'
Roll
centre
Fig. 4.9 Roll moment.
Roll geometry 63
is defined by the geometry of the suspension system, as will
be discussed later in this chapter. The roll moment which
must be resisted by the suspension springs (and the anti-roll
bars, when these are fitted) is equal to F x h, where F is the
centrifugal force and h is the distance between the sprung
mass centroid and the roll centre. For a given spring rate, the
roll resistance will be greater and the roll angle less if h is
reduced.
x
I
I--~
FIG Wheel R
rP==_=-1--:- .- - or -----
-==---
0 0~~ ~
_ -' _
I
Xl --- --
v --
Fig. 4.10 Determining the roll centre.
64 Suspension geometry
will move at right angles to line PCL to a new position Q. At
the same time the decrease in load R2 will cause the contact
point V of wheel R to move to position V. If we rotate the
body in an anti-clockwise direction until the wheel footprints
have returned to their original positions P and V, the only
point about which rotation can occur within the constraints
that wheel L must move about centre CL and wheel R move
about centre CR is seen to be point Mo. The position of Mo is
found by the intersection of lines drawn between CL and P
and C R and U. In some cases, notably the popular
MacPherson strut suspension, the roll centre is not at a
constant height above ground and moves slightly as the roll
increases.
• - - • - - - --11-1<>--1-
-- -- - -r-!-'~--t--~::t-<>I
. - - -- - - ...,....,....,~...,......,.....,.-.,rl......."....,~...,...-r-~~/
M"
Fig. 4.12 Roll centre with double-wishbone suspension of equal-
length parallel links.
Roll geometry 65
Mo
Fig. 4.13 Roll centre with vertical sliding pillars.
length (see Fig. 4.12). Since the instantaneous link centres are
at infinity, the roll centre must be at ground level. The second
system is the old Porsche trailing link system using
equal-length parallel links as used on th~ familiar VW Beetle.
This system is shown in Fig. 5.10 in the next chapter. The
third is the sliding pillar suspension with vertical pillars, as
shown schematically in Fig. 4.13. When the sliding pillar is
set at a swivel angle (see Fig. 5.9 in the next chapter), the roll
centre falls below ground centre. The position is defined by
extending the pillar centre line to intersect the ground plane,
then by drawing a perpendicular line to intersect the vertical
centre line of the car.
The de Dion suspension system separates the two functions
normally performed by a live rear axle. The final drive unit is
mounted on the chassis (or body) and is, therefore, part of
the sprung mass. Open halfshafts with universal joints at
both ends take the drive to the wheels. The de Dion tube
which connects the two wheels and maintains them in perfect
parallelism is always located laterally, either by a vertical
slide, as shown in Fig. 4.14, or by other effective means.
With effective lateral location, the body is constrained to roll
about the centre point of the de Dion tube.
The swing axle is now obsolescent. It was used by
Profes.sor Porsche for the rear suspension on many of his
9'
SWIVEL
ANGLE
I'
CAMBER
ANGLE
FD o
M, ·- .---SentrO
- _ _ _ R OI(
. ~ . ox's
G-=--- ·-.....;...."O"'is
Fron l roll
id .
---.:... .
rOUi'>(j ·:-:-....... _ _ .
Pion c :::-..- ___
entre_/,:--"
- ._ .-
@
~ ,
M,
. ___ ~ l /
F
me 7~
~ centre
Reor roll
Therefore
(4.2)
72 Suspension geometry
4.3.1 Single wheel bounce (with anti-roll bar)
Applying the above to the case of single wheel bounce, we
can take the main spring rate at the wheel as k and the
anti-roll bar rate, also measured at the wheel, as k.. The
single wheel bounce rate ke is, therefore, given by:
ke = k + kka k . (4.3)
a+ k
A good example to illustrate how the anti-roll bar can
influence the single wheel bounce rate is the Porsche 928,
which is the subject of a design analysis in Chapter 12. The
Porsche 928 is provided with a strong anti-roll bar at the
front and a relatively weak one at the rear. The appropriate
data are as follows:
Front Rear
Wheel rate (kN/m) 18.63 22.55
Anti-roll bar rate (at wheeI)(kN/m) 83.4 10.2.
Single wheel rates in bounce
Front 83.4 x 18.63
18 . 63 + ----,-----,--~
=
83.4 + 18.63
= 33.86 kN/m
Fig. 4.22 Anti-dive system, using leading arms at the front and
trailing arms at the rear.
pL 11-pi L
pL 11-piL
o
y
4.4.1 Anti-squat
Anti-squat is resistance to squatting down at the rear end, and
as we may see in its more spectacular forms on the drag
strips it can be resisted in the limit by the provision of rollers
at the rear extremities of the body. Fortunately for the
designer of more ordinary vehicles, the provision of 'leading
arm effect' at the front and 'trailing arm effect' at the rear
works well in the reverse direction. Some compromise is
necessary, since an anti-dive geometry carefully calculated to
match a particular front/rear brake distribution will seldom
give a perfect correction for anti-squat.
In Formula 1 racing the problems of anti-dive and
anti-squat became acute in 1979, when the majority of
Formula 1 cars adopted 'ground-effect' body designs that
pulled the car down on the road by the vacuum created on
the underside of large side pods. Sliding skirts were devised
to seal the body sides to prevent leakage of higher-pressure
air to destroy this vacuum. These skirts actually touched the
road surface and often made grooves in the tarmacadam
when the suspension exceeded the available movement of the
skirt in its slides.
The most successful designs of the 1979 racing season were
those that achieved the closest approach to a no-roll
suspension with negligible dive and squat. The latest type T4
76 Suspension geometry
version of the Ferrari 312 uses heavy Bowden cables running
from front to rear to give a coupled front/rear suspension
system designed to reduce dive and squat. As I write, the
outcome of the experiment is unresolved, but at least they
have won the Constructors' Championship for the year.
5
Conventional Systems
5.1.1 Shimmy
The driver could feel the onset of shimmy when the steering
wheel began to vibrate rapidly from side to side. This,
precisely, was what was happening to the front wheels.
Accessory manufacturers made steering dampers designed to
reduce the amount of shimmy. In racing and sports cars the
designer provided very stiff springs to reduce the angular
deflection of the front axle, since this angular deflection of the
axle is the source of shimmy.
Any rotating mass, such as a wheel and tyre, exhibits
gyroscopic effects. Thus, if a spinning wheel is subjected to
78 Conventional systems
an angular displacement, a force is produced tending to move
the axis of rotation in a direction at right angles to the original
displacement. The reader can demonstrate this phenomenon
using an ordinary bicycle, preferably one with large-diameter
wheels. He should stand astride the bicycle and lift the front
wheel clear of the ground. Then holding the cycle in this
position with one hand only on the handle bars, spin the
front wheel as fast as possible in the normal forward
direction. Finally, with the wheel spinning fast, he should
grasp the handle bars with both hands and lean the cycle to
the right. The gyroscopic torque will be felt as a distinct
reaction at the handle bars as the wheel tries to tum to the
right. If the wheel is spun in the reverse direction, the
steering reaction will be to the left.
This serves to illustrate the gyroscopic torque that twitches
the front wheels to left or right when passing over a bump or
a pothole in the road (see Fig. 5.1). With inadequate damping
of the front suspension (a characteristic of early vehicles), the
inertia of the heavy front axle and the masses of the wheels at
each end induces an oscillation at the natural frequency of the
springs. As long as these oscillations continue, with one
wheel rising as the other falls, the gyroscopic torque will
produce the steering oscillations of shimmy.
5.1.2 Tramp
With a beam axle tramp can occur at front or rear. Even when
shimmy is eliminated a beam axle can still develop tramp,
which is a transverse rocking action at spring frequency that
can persist for many seconds. The most common cause of
oc.=Angu/ar def'/ection
of' axis of' wheel
Fig. 5.1
The live rear axle 79
Braking
torque·
5.2.1 Disadvantages
The live rear axle is relatively heavy in comparison with an
independently suspended one. The outer casing is
constructed from malleable castings and/or steel pressings and
contains the final drive crown wheel and pinion, the
differential casing and its gears, the halfshafts and all the
necessary bearings to support the moving parts. The
customary drum brakes, brake shoes, backplate, wheel hubs,
wheels and tyres all contribute to a fonnidable unsprung
mass.
The unsprung mass is approximately twice that of a good
design of independent rear suspension. This in itself is a
serious disadvantage as was demonstrated in Chapter 2. A
car with a low ratio of unsprung-to-sprung mass (m/M) will
maintain tyre/road contact on a rough road much more
effectively than one with a high ratio. In Chapter 2 it was also
shown that a high m/M ratio is also associated with a hard
ride.
Apart from the low cost of the live rear axle, it does
represent a very robust design and with the help of modem
techniques of longitudinal and transverse location satisfactory
rear suspension can be given for a low-cost front engine/rear
drive family car. The word 'satisfactory' is used advisedly,
since the high standards established by a good design of IRS
cannot be matched.
Torque
~ube
Telescopic damper
plus auxiliary spring
Fig. 5.6 The Rover 3500 live rear axle with short torque tube.
84 Conventional systems
chosen for the Rover 3500 as shown in Fig. 5.6. The coil
springs are placed immediately above the axle tubes. Forward
of each tube are combination spring/damper units. These are
located on forward-facing brackets at points on the lines
drawn between the torque tube universal joint and the wheel
centres. In this position the spring/damper units can control
body roll as well as bump and rebound. Lateral location is
given by a Watt's linkage.
\ \. . -. - ~
.-" --~ --
: 1
...I
I· I
I
.,-v--r-- ~ =='.-=.".-=-
I.
i *I
, .'
--.-_. /"' " Roll
centre
9 2 3 8 7
1-6 0°
1-4 + SO
z +10°
~
1·2
III
0
L
.2 1·0
01
C
·c
III
c
L 0·8
0
u
0·6
0·4
0·2
o 2 4 6 8 10
Slip angle (degrees)
Fig. 5.11 Influence of camber angle on cornering force.
~.----
ZERO CAMBER
X
/'
, '
.....
~1IlllITl=_
I I
with the effective brake. With hard braking this could result
in a spin. With negative offset the steering pull is in the
opposite direction, giving a better . chance of straight-line
braking under partial brake failure. A very useful bonus is a
reduction in the danger of spinning when a front tyre blows
out at speed.
B~ftRoll centre
SWING-AXLE
~ Roll centre
TRAILING LI NK
r
--
SEMI-TRAILING LINK
__ PLAN
~
I I ...............
::
~-=.::,.-=.-=.~ _ L - ____ ~
~ I 0
tW-
Effective swing-orm length I I
~oll
END ELEVATION
Effective
__ _ centresWing-arm
Fig. 5.20 Camber angle change under roll with swing axle, trailing
link and semi-trailing link suspensions.
Independent rear suspension. 103
~·~·,tm
~.II
'///
!O!315rh
I I
//'---\\\ ---t-.
I ~'"
n
~ I O'64m
I
O'65m ,~ . ! \ i, I O·965m
e=25_1_5_ _ rRear~alls _ \ I .
\ I
\~
~
Car
! Plan
1·42m
i
--~~----r-----
Axis and rear whLI centre line
, --------------- - '
Rear elevation
Fig. 5.21 Typical semi-trailing arm layout.
/1~
References
~
__
(a)
Turning centre
t=~ --'-40
H4'
"
___ . ______-----------
~~ Ackermann
centre
1\
Sr
( b)
Tr
R·I
-Rr
- Rf
(e)
Rj
Fig. 6.1 Vector diagrams for cornering RWD and FWD medium-
size saloon.
l . _________
centre0
~~-~- -~~-~-~~:=-=---~
~-- Ackermann
i\ (a) centre
s,
5;"-----
---
(b)
Fig. 6.3 (a) and (b): Vector diagram for cornering racing car.
725 X 30 2
50
= 13050 N.
This is a centrifugal force of about 1.8g.
112 Road-holding
In Fig. 6.3(b) the broken lines e', T' and S~ show how the
driver can lift his throttle foot, reduce the torque from the rear
wheels and change the turning centre. This, of course, is a
steering action. The changes in Sf and Sr are negligible for a
large change in direction, giving the car good stability in a
comer.
/ --- ----
'1- ---, - If- J.--'t1--;r---
, ____ !..." ~ )0",
______ • ~
,- - - ,~
,.,.
__ -
-~.
\ .. I. \
--
1
'--I) I -\-,_; ...... _ - _ .... ', 1 I I
1 I: --- )..l. I~ __ J
: ::,' r---"'_ ,.- ... ~I-V' ,----.I
I
-- --.l- --' -,: -:. :- ,. .).: --
I J "... ,
l
,L _ _ _ .!j'" __ .. ..:.-:, ... --~.!'-.-'_~-;---
....
I I,~'
'T,' I
, _ _ _ _I 1_ I
,
'''-.,
.... --- ----- /
-
(a) HIGH POLAR MOMENT OF INERTIA
, ---, \
~
\ _ \ - ..... _ _ ,
:- -=-=-== .... \
,
I
. -
---=L
-- __ "
' I
.---+
"-. --- t - - - - I
:-- - - -! -- -M 1 1
1-----1 ' __I L _____'
""~ s ---.!L
- 11 + 12 . (6.2)
v 7
w = r= 10 = 0.7 rad/s.
L s = 2S00 x 0. 7 = 620 N.
2.8
6.7 Roll-steer
When a wheel is deflected from its normal axis of rotation by
body roll, it can influence the car's direction. When both
wheels at the front or both wheels at the rear are deflected by
roll, the roll-steer effect can be quite pronounced. It all
depends upon the suspension geometry.
An example of roll-steer can be illustrated by the simple
example of a dead rear axle (see Fig. 6.6) as used on many
FWD cars in which the trailing links slope upwards to the
rear. As the body rolls, the more heavily loaded outer wheel
moves upwards and fonvards while the inner wheel moves
downwards and backwards. This produces understeer. In
practice the link pivot centre would be only slightly below
wheel centre height, probably no more than 50 mm. The
same technique can be used with a semi-trailing link
suspension or with any independent suspension design using
longitudinal links to control the fore and aft location of the
wheels. It can be applied to both front or rear suspensions.
Excessive roll oversteer can make a car 'twitchy' on an
uneven road surface. A memorable example was the original
Mercedes-Benz 300SL with double-wishbone suspension at
the front and swing axle suspension at the rear. The low roll
centre at the front, the very high roll centre at the rear, plus
the variable cornering power produced by single wheel bump
and rebound at the rear as the rear wheels alternate between
positive and negative camber, all contribute to a lack of
stability of speed. (I myself remember this tendency to weave
from side to side at high speed!) Daimler-Benz removed this
inherent instability when they introduced the low-pivot
swing axle design.
118 Road-holding
, I
Toe-ou~ \
for \
------- --
undersffier~
---
----- ---- --
~----
_ - .. 0
Toe-our
for ~
overs~eer
. --- --- - - - - --- --- --- --- -'---.
-- -
I ._
. .J-"'---
Burnp"a\~2.---- I
J...---
,',
I
-.-
.--
I
,
~' , I
",,' I I
Speed
Fig. 7.3 Typical flow characteristic curves given by bleed flow and
linear valve flow.
122 Dampers
body, the cylinder to the unsprung mass. It is easy to see that
the use of a viscous fluid in the damper will give a damping
force that is 'velocity-dependent', i.e. F ex: V n • The value of the
index n will depend upon the design of the spring-loaded
valve. At very low vertical wheel velocities the valve is
designed to remain closed, all oil flow being through the
bleed orifices. These are small enough to give streamline
flow. The modern 'ride-control' or 'linear' valve is
spring-loaded, as shown in Fig. 7.2. By changes in spring
strength and the diameter and port dimensions of the valve
orifice, a wide range of resistance/flow curves are possible. As
the linear valve opens under high fluid flows, a point is
eventually reached when the valve disk is lifted so high that
the flow is controlled by the diameter of the valve orifice.
This gives a gradually increasing rate of rise in resistance, as
seen in the rebound valving curve of Fig. 7.3.
Most damper designers provide different valves for bump
and rebound flow, the resistance usually being much less
in bump than in rebound. Typical characteristic resistance
curves are given in Fig. 7.3.
Seals
Rod
Cy li nder
Reservo ir
Non-return
V al ve
Piston
Cal ibrated
Cha nne l s
By·pass
Valve
Foot
Valve
• ". -, .1
'.
. ~ . .
00 0
o
o
o
o 0
o 0
o
0.
Cc =.f(4kIM)
C >!c4kIM)
,
11/
>
C=J4kM
e
(7.1)
7.5.2 Resonance
We have already referred to the phenomenon of resonance in
Chapter 2. As shown in Fig. 2.6, with a complete lack of
damping the effect of a regular wave formation in the road
surface could be quite shattering when the forcing frequency
ir coincides with the natural spring frequency in. Curve 1
demonstrates this. Curve 2 with a maximum acceleration at
resonance of about 0.7g, has a value of CjC c = 0.15. Curve 3
is for a value of CIC c = 0.25 and Curve 4 is for CIC c = 0.65.
Curve 3 has been labelled 'well-damped', but these curves
are based on a damping function directly proportional to
vertical wheel velocity. Damper designers have accumulated a
wealth of practical experience in the design of ride-control
valves that gives a finished product giving good damping
over a very wide range of frli n values.
References
M= 400 kg
TIV
250mm
11----;--11
9·0
8'0
~ 7·0
z
~
Q/ 6'0
~
.2
] 5'0
I..
~
4·0
3·0
2·0
1·0
~- - -
Fig. 8.3
.1
BEA.O RING I
I STEEL BEA.O ,/
TO AIR SPRING
AIR INLET
FROM
STORAGE
TANK
I
10
I
9
I
8
x 6
c ,/
/
~
I 5
~
'" f
/ B"" ""
1
~. V .... "'"~"'" .... ...-
-.......:: ~. 4..: ~-
o 1 1 1 4 5 6 7
STROKE-in.
EXTENDED COMPRESSED
'"~ ~
E
~'OO
>-
u
z
w
::I
:Z90
---r----
:f
"fso
Z r--..
'"
70 0 10 20 30 4OxI0 2
SURGE TANK VOLUME (inl)
t
"'EBOUND
SPRING STROKE
BUMP
Fig. 8.9 Dynamic load curves using pedestals of different proffies.
138 Pneumatic suspensions
8.3.2 Single-convolution air springs
With the automobile market in view the Firestone Co.
developed a rubber and canvas bellows with only one
convolution. To vary the spring rate characteristics, curved or
tapered pedestals were developed to give an effective piston
area that changed throughout the stroke. Fig. 8.8 shows two
such pedestals, and Fig. 8.9 shows how a change in pedestal
profile can influence the load/deflection curve. With pedestal
2 soft suspension will be given for small deflection in both
bump and rebound (curve BC). For greater deflections in
bump the spring operates over the rising section (CD). In
rebound the spring operates over a more modest rising rate
curve (BA).
RH front suspension
cylinder and sphere
R H rear suspension
cylinder and sphere
LH rear suspension
,,====~CY~lin;d~er and sphere
Rearheight
corrector
Gear selector
return
~-----~--
Pressure /Steering
regulator return
return
Nitrogen
ot the initiol
inflation
pressure
Nitrogen ot
the pressure
of fluid in use
overfIOW_~===1
return ... Push rod
Venrto
atmosphere
Front
Rear
Disk tance
volves Calibrated spacer
bypass
Fig. 9.4 Front and rear suspension cylinders with enlarged cross
section of damper valve body.
The Citroen suspension 147
synthetic rubber diaphragm at the initial charge pressure. The
gas is dry nitrogen. On the right the accumulator is fully
charged with hydraulic fluid at the regulator cutout pressure.
The pump is supplied with fluid from the fluid reservoir
which is not only a reservoir, but a filtration and degassing
tank into which all fluid is returned from the various
operating units, such as the suspension units, gear selector,
etc.
(a)
Rubber
_~;...,..~~~---t:l-l,L..~_?" diaphragms
Disks ~====~;;"""I;:S:l::;;;::J~~
Springs ..",;::::::::,,-~s-1/T
Suspension cylinders
( b)
(d )
Fig. 9.7 Operation of the height corrector. (a), movement from
cutoff to exhaust position: When the slide valve is moved, i.e. when it
moves its position from cutoff, the disk valve in chamber C is held
on its seating by a return spring, thus closing the clear passage. The
disk valve in chamber D is lifted off its seating by the shoulder on
the slide valve, thus opening the free passage. The fluid in chamber
C is therefore obliged to pass through the dashpot which slows
down the fluid movement, which in turn slows down the movement
of the slide valve. Thus the slide valve will not move to the exhaust
position unless there is a positive effort on it for a certain period of
time. No correction occurs for rapid wheel movements; (b),
movement from exhaust to cutoff position: When the slide valve is
returned to the cutoff position, the fluid in chamber D can this time
use the clear passage and return to chamber C, lifting the disk valve
against its return spring. Return is therefore rapid. As soon as the
slide valve returns to the cutoff position, the disk valve in chamber
D closes the passage again, stopping the slide valve over-running the
cutoff position and avoiding a second correction; (c), movement from
cutoff to inlet position: When the slide valve is moved, the disk
valve in chamber D is held against its seating by its spring, closing
The Citroen suspension 151
9.2.5 The function of the anti-roll bar in height control
Rotation of the anti-roll bar is used to control the movement
of the slide valve. We must first consider how rotation of the
anti-roll bar occurs. During roll one arm of the anti-roll bar is
forced upwards, the other arm downwards. Since these
movements are equal and opposite the bar is placed under
torsion, but no actual rotation occurs relative to the
supporting bushes carried by the body. Simple single wheel
movement in bump or rebound will produce some torsion
and a little rotation. However, if both front wheels or both
rear wheels move upwards or downwards in unison, the
anti-roll bar arms on both sides move in the same direction.
This produces rotation of the anti-roll bar and is the basis of
the Citroen system of height control.
The complete control mechanism is shown in Fig. 9.S. The
height control rod is clamped to the anti-roll bar and passes,
parallel to the bar, to a lever behind the corrector valve. This
lever is pivoted at its lower extremity. Rotation of the anti-roll
"bar moves the corrector valve slide inwards or outwards, as
described in Fig. 9.7. Rapid wheel movements, i.e. under the
action of single- or double-wheel bump or rebound, have no
effect on the corrector valve, since the movement of fluid
between chambers C and 0 is restricted by the 'restricted
passage' shown in Fig. 9.6. Height correction can only occur
after a delay of several seconds.
An over-riding manual control, located near the driver,
allows him to increase the ride height when travelling over a
rough terrain or to clear snowdrifts. It is also used when
changing a wheel.
the clear passage. The disk valve in chamber C is lifted off its seating
by the shoulder on the slide valve, thus opening the clear passage.
Liquid in chamber D therefore has to pass through the dashpot. As
for operation (a), there will be no movement of the slid~ valve until
a certain designed time period has occurred; (d), movement from
inlet to cutoff position: When the slide valve is returned to the cutoff
position, the fluid in chamber C can this time use the clear passage
and return to chamber D, lifting the disk valve off its seating. As in
operation (b), the return is rapid, As soon as the cutoff position is
reached, the disk valve in chamber C reseats. This stops any
over-run of the slide valve and prevents a second correction.
152 Hydropneumatic suspension
Suspension cylinder _ _~....-
Front 'halfaxle'
mounted on chassis
Height corrector
~ Increasing load
[::>i> Decreasing load
Overflow returns
Control rod
Pivot point
~;>.--_ _ Suspension
{,.--_-li\ sphere
Suspension cylinder
Bump as rebound
fers
Height
corrector
Ant i-roll
bar
9.2.11 Anti-squat
It would be very odd indeed if Citroen introduced all this
complex hydraulics into the suspension design in order to
provide a constant ride height and then completely ignored
the longitudinal dips and dives induced by braking and
acceleration.
Anti-dive geometry in general was discussed in Chapter 4.
)F $
The centroid in the Citroen is well forward, with a one-up
weight distribution of 62/38 and a full-load distribution of
$£':-: ___= __
Centroid
(T (f"";'OdOn)
, - -$,;:... I~
Centroid
(one-up) (fully laden)
Of Fi=r2 Pr
,~-~J
p. H
-- 1I==-'--+-
f
Braking
Fig. 9.12 Anti-squat and anti-dive geometry on GS model.
The Citroen suspension 157
54/46. The force diagram in Fig. 9.12 is a compromise based
on a weight distribution of 58/42.
To achieve perfect anti-squat geometry, the decreased loads
on the front footprints and the increased loads at the rear, as
shown in the upper diagram in Fig. 9.12, must be balanced
by a vertical downthrust on the suspension arms at the front.
The tractive force F from the front wheels produces an
acceleration IX. Approximately 10% of this tractive force is
expended in accelerating the unsprung mass and about 90%
in accelerating the sprung mass M s:
0.9F = F j = MslX.
F j is the inertia force acting on the sprung mass. This tends to
rotate the sprung mass about its suspension anchorages. This
decreases the load on the front wheels by AL and increases
the rear wheel loads by the same amount:
H H
AL=F.-=MIX-
IE s E .
9.2.12 Anti-dive
With a trailing arm suspension at the rear and a wishbone
system at the front, angled to give an effective leading arm
action, the Citroen GS has an ideal geometry for anti-dive
correction. As shown in the lower diagram in Fig. 9.12, the
values of AL now become an increase in load at the front and
a decrease at the rear. The braking forces applied to the
158 Hydropneumatic suspension
footprints B f at the front and Brat the rear, produce reactions
at the suspension points P f upwards at the front and P r
downwards at the rear. These forces, with careful design,
will balance the couple AL x E.
The Citroen GS has a brake pad area at the front of
146 cm 2 and at the rear of 68 cm 2 • This suggests a front/rear
braking effort of 68/32. The situation is more complicated,
however, since the front and rear brake cylinders do not
operate at the same pressures or even at constant pressures.
The front brakes always operate at normal full-circuit
hydraulic pressure (circa 160 bar). The rear brakes operate at
rear suspension pressure, which is nominally 35 bar. When
fully laden the rear suspension pressure increases to as much
as 65 bar, but this is still well below the operating pressure at
the front. It is obvious that the front braking bias is so high
under all load conditions that the front wheels will always be
the first to lock. This is an established prerequisite for safe
braking in the wet. The lower diagram in Fig. 9.12 is given
purely to show the principles of anti-dive geometry, using
leading and trailing arms. We do not have sufficient data to
put concrete values to the Citroen design.
sure line
Control pressure line
Fig. 9.13 Hydraulic circuit for hydropneumatic suspension on Mercedes-Benz 450 SEL 6.9.
c::
-
o
'"c::
~
8-
E
o
u
162 Hydropneumatic suspension
Damper pi~ton
Pressure
Piston rod ----H;;.. line
Pressure cylinder
Gas-filled
pressure reservoir
F F
5 = 5 xs
e 5 +s .
Sf Sf Sf Sf
Damper
valve
Tapered
cylinder
j)iaphra~m
NORMAL BUMP
Rubber spr;n~
Damper hleed-.-;;="'="nll
Tapered piston
10.3.1 Interconnection
The Hydrolastic spring elements on the same side of the car
are interconnected by small-bore piping. The pipe connection
'cannot be seen in Fig. 10.4 but is made at a high point on
the upper surface. Fig. 10.3 illustrates how single wheel
movements were controlled by both front and rear springs on
the same side working in series. With very large-bore piping
this would have effectively halved the spring rate in
comparison with the use of isolated units. In practice the use
of small-bore piping modified the true series operation to an
intermediate effective spring rate. The damper design also
modified the overall behaviour. Under roll, however, the two
springs on the same side worked in parallel and, as we
know, springs working in parallel are additive.
The Moulton Co. have a long experience of the design of
rubber suspension units and they were able to provide a
rubber doughnutlike spring (see Fig. 10.3) which had been
carefully designed to give a rising rate suspension. With very
large-bore pipes the only resistance to movement in pitch
would be the resistance to flow designed into the damper
valves. With small-bore pipes the pitch frequency must
increase for larger movements. From experience, I can state
that pitch control on this vehicle was very good and for such
a small car the ride was excellent. Today some small cars with
conventional springing have an inferior ride.
170 Interconnected and no-roll suspensions
- .
HYDRAGAS
...
//~~"\16_ Lorge Small
area area
PI TCH
Gas
HYDRAGAS
BOUNCE
Fig. 10.5 The Hydragas principle.
Rates. The makeup of the bounce, roll and pitch rates, using
the terms used by Moulton Developments Ltd, is as follows:
Hydraulics* - that due to the compression of the gas acting
through the liquid (at the same pressure) upon the
diaphragm.
Taper - that due to the pressure of the liquid acting upon
the changing area of the diaphragm, as it is actuated by the
suspension arm.
Parasitic - that due to rubber bushings in the complete
suspension system.
Drop angle - that due to change of leverage with stroke.
Essentially in the Hydragas system the pitch rate is much
lower than in bounce; the bounce rate and the roll rate are
identical since no anti-roll bars are used. As shown in Fig.
10.7, the pitch rate is made up of taper, parasitic and drop
Rear
1550 Nm (40"/0)
~
1~8°
il
_.-tI
Centroid of sprung mass
4400 N
I 1600 N
j .- -
0, =--:-------~-~ ~O,
Roll centre
2935 N (28"/0) 1070 N (10"/0)
Fig. 10.8 Analysis of tyre cornering forces and roll moment at front
of Lotus 79.
176 Interconnected and no-roll suspensions
00
I
2330 Nm (60%)
~
1f- 20
I Cen~roid
.1
of sprung mass
I II
'.
---~
I
I I
6450 N
~~~-.r--4~--~====~~_~_~~~~~~
il __ ~
Roll centre 0,
4670 N (45%) 1780 N (17%)
Fig. 10.9 Analysis of tyre cornering forces and roll moment at rear
of Lotus 79.
/
I
/
/
RCl
Ground
Fig. 10.10 Trebron double roll-centre suspension applied to racing car - static condition.
j t
'b
I
'"
.!l
'b 0
f"
ll=-~
Jill
Hydrauli c Main
strut Rea r roll
c ontrol
Gas sp ring
an d
control valve control valve damper
Fig. 10.12 Line drawing of AP fully stabilized suspension.
182 Interconnected and no-roll suspensions
As in the Citroen system, each wheel is provided with a
gas spring and damper strut which is connected to a
suspension arm. The level of the car body under the action of
roll or pitch is corrected by very rapid additions or extractions
of hydraulic fluid from the space between the gas spring and
the damper piston. The signal to add or extract fluid is given
by a three-port valve. What the maker calls a 'pendulous
mass' is used to sense any change in level (see Fig. 10.13).
This mass is supported on a spring with a small hydraulic
damper in parallel. For any application the
mass-spring-damper unit must be specially tailored to match
the suspension system. In effect, it is a replica of the car's
suspension system in miniature. Thus under single wheel
bump the action is as follows: upward movement of the
suspension arm compresses the gas spring and creates an
upward force to lift the body. The pivot of the offset
pendulum also moves upwards, since it is attached to the
body. If the pendulum did not move, the spool of the
three-port valve would move to the left and extract fluid
from the suspension leg. The pendulum-spring-damper unit,
having been designed to behave as a replica of the
suspension system, will produce an upward movement of the
pendulous mass at an identical velocity to that of the body.
~~~~jjlsupporting
t~~~~~ spr ing
\:jg~~:::;;:;::~ Pendulous
pump
References
Today, very few cars are designed by one man. Even cars
that are made in half-dozen batches like Formula 1 racing cars
have a specialist like Keith Duckworth to design the engine.
Again, Hewland Engineering Ltd design the transaxle and an
overall chassis specialist like Derek Gardner designs the rest.
When a mass-produced car such as the Ford Fiesta is
designed, relatively large specialist teams are involved with a
project leader to co-ordinate every part of the project with the
188 A small FWD saloon car: Fora Fiesta S
whole. Inevitably, committee efforts are slow but the vast
inveshnent involved encourages caution. The executive
engineer in charge of chassis design on the Fiesta project was
Tony Piilmer and his original plan for a wishbone suspension
system at the front using torsion bars was changed to the
MacPherson strut system, a design used on several earlier
British Ford cars.
Wheelbase, 2.286 m
Front track, 1.334 m
Rear track, 1.321 m
Kerb weight, 727 kg
Laden weight (4-up), 1040 kg
Sprung mass (4-up) (estimated), 945 kg
Front/rear weight distribution (4-up), 54/46
Tyre size, 145-12
Front spring rate, at wheel, 21.7 kN/m
192 A small FWD saloon car: Ford Fiesta S
Rear spring rate, * at wheel, 25.0 kN/m
Rear roll bar rate, at wheel, 3.3 kN/m
Total wheel travel, front, 143mm
Total wheel travel, rear, 166mm
Wheel travel in bump, front, 62mm
Wheel travel in bump, rear, 84mm
Height of roll centre above ground, front, 185mm
Height of roll centre above ground, rear, 191 mm
Roll stiffness, front, 296 Nm/deg
Roll stiffness, rear, 210 Nm/deg
Roll angle at a lateral acceleration of O. 5g, 3° 54'
Camber angle, static, at kerb weight, 1 ° 57' positive
Camber angle, static, laden, 0° 52' positive
Toe setting, 2.5 mm, toe-out
Kingpin inclination, 15° 6'.
185 + 191
Mean roll centre height = 2 = 188 mm.
• This rate is based on both rear wheels in bump and rebound. Single
wheel rate is 12% higher.
The Fiesta suspension 193
Fig. 11.6
Rear Front
conjugate conjugate
point point
--~~--_Il--~~.~!
..I. s~
Fig. 11.7
Upper and lower wishbones are used at the front. At the rear
upper and lower transverse links are used with the addition
of a special patented flexibly mounted trailing link known as
the Weissach-Axle. The suspension geometry gives 30%
anti-dive at the front and 50% at the rear. The rear
suspension layout also gives 70% anti-squat. Compromise is
essential in these factors, especially when designing a sports
car, if adverse changes in toe-in, camber ~nd castor angles are
to be avoided. The suspension engineers at Weissach, the
Porsche research centre, admit that they arrived at the chosen
geometry by trial and error. With a new untried design one
can hardly expect the computer to give a reliable answer.
On the front suspension an alloy casting is used to provide
a conventional upper wishbone. An even more robust alloy
casting is used for the transverse lower arm. These
components are particularly robust since they are subjected to
very high forces during braking, acceleration and cornering.
'By comparison the suspension components on the Fiesta
are quite flimsy. The Porsche 928, however, has a
power-to-weight ratio of 115 kW/tonne (160 b.h.p./ton) and
the Pirelli P7 225/50 VR16 tyres, fitted as standard, endow
The Porsche 928 suspension 201
the car with phenomenal cornering accelerations. The use of
tyres with such a low-profile ratio does, however, involve the
penalty of a suspension geometry that allows only small
changes in wheel camber. At the front the variation from full
bump to full rebound is only about 1.5°. At the rear it is less
satisfactory since it approaches 4 degrees.
The components of the rear suspension can be seen in Fig.
12.3, which is a photograph of the rear suspension
cross-member and the suspension components taken from the
front. Each wheel hub carrier is located by a simple
transverse upper link and a more complex lower link which
is, in effect, an integrated two-part link. The transverse
member is a relatively thin steel plate, very stiff under pure
vertical loads, but flexible in the fore-and-aft direction. The
trailing link is attached to an articulated double-pivot, which
lies approximately in line with the inner pivot (the pivot of
the steel plate) to make an angle of approximately 25° with
the centre line of the car. The Weissach-Axle is, therefore, a
semi-trailing link suspension with a variable axis. The
dangers of excessive amounts of compliance in the rubber
bushes used in suspension links are well known to
suspension engineers. In racing car design, where small
variations in suspension geometry cannot be tolerated, rubber
bushes are never used. Drivers of ordinary passenger vehicles
do not expect to suffer shocks, vibrations or noise transmitted
directly from the suspension attachments. Isolation, within
Decelerating
(a)
Deceferating
( b) (c)
Fig. 12.4 (a), Porsche 928 conventional suspension toe-out due to
elasticity. When decelerating or braking, the wheels in a
conventional suspension toe-out at a angle owing to force P. The
reasons are the rubber bearings in the suspension required for noise
absorption. (b)-(c), Weissach-Axle - toe-in due to kinematics. A
kinematic effect changes the angle towards toe-in when decelerating
or braking. For reasons of nOIse absorption, rubber bearings are
featured, but toe-out is immediately balanced by the kinematic effect
a = p. Thus dangerous side movements are prevented in curves, an
important characteristic.
The Porsche 928 suspension 203
the limits of current technology, is therefore the order of the
day, even on a sports car.
Cunning old Americans say 'If you can't lick 'em, join 'em'.
This is the philosophy of the Weissach-Axle. Rubber
mountings are used on the Porsche 928 to modify suspension
behaviour under traction, free roll and deceleration to give an
automatic correction to the amount of rear wheel toe-in. Even
when cornering near the limit, the fine tuning of the variation
in toe-in calls for negligible steering correction from the driver
when forced to decelerate.
With greatly exaggerated angles, the Weissach-Axle
principle is explained in Fig. 12.4. On a conventional
semi-trailing link rear suspension deceleration (and to a
greater extent, braking) deflects the rubber-mounted pivots
to give a toe-out angle to both rear wheels. This is an
oversteering effect. Since braking also robs the rear tyres of
effective cornering force and increases the rear slip angles (see
the Circle of Forces, Fig. 1.17), these two oversteering effects
in combination can sometimes give the driver a difficult
handling problem when decelerating in a corner. To the
professional test driver, this is often termed liftoff tuck-in.
When very rapid unwinding of the steering wheel is
necessary to correct this oversteer, the car can hardly be
called 'inherently stable'.
The front pivot of the Weissach-Axle is shown
schematically in Fig. 12.4. It consists of two bushes, a
conventional rubber bush on a rod end and a larger rubber
bush, 100 mm long and 50 mm diameter, to which the rod
end is fixed. Only very small changes in the amount of toe-in
are permitted by this double-jointed pivot system. The static
toe-in is 20', decreasing under traction, but never reaching a
toe-out angle. Under severe braking, the toe-in will increase
but will seldom reach more than 50'. These are very small
angular changes, but sufficient to give vicefree cornering
behaviour.
The following data have been supplied by Porsche:
Type 928 chassis
Wheelbase, 2.5m
Track, front, 1.545 m
Track, rear, 1.514 m
204 A high-performance sports car: Porsche 928
Kerb weight, 1450 kg
Permissible total weight, 1870 kg
Weight distribution, one-up, 49/51
Weight distribution, two-up plus luggage, 47/53
4754 = 1. 7° = 1 42'.
0
2770
It is interesting to compare this with the Ford Fiesta S in
Chapter 10 which rolls through 3°54' under 0.5 gravity
centrifugal acceleration.
The camber changes on the inner rear wheel and the two
front wheels will be negligible. Thus by the provision of good
Natural frequencies and double conjugate points 207
roll resistance, Porsche have met this essential requirement for
a car fitted with 50 series tyres.
Therefore
K2 = 1.5306
c2 = a x b = 1.548
x = a -II = 0.094 m
x2 = 0.0088.
From this data we can obtain the following values:
Conjugate Points (one-up)
Front conjugate point distance behind centroid s = 1. 382 m
Rear conjugate point distance in front of centroid r = 1.103 m.
* For the type 924 (the 2-litre model) Cranfield Institute measured a value
of 0.95.
208 A high-performance sports car: Porsche 928
Natural frequencies (one-up)
Ff = 1.18 Hz
Fr = 1.28 Hz
F pitch = 1.24 Hz.
12.4 Comments
The normal wheel rate at the front is 18.63 kN/m and the
anti-roll bar rate is 83.4 kN/m. The additional equivalent
spring rate is therefore:
The total spring rate for single wheel bump at the front is
therefore 18.63 + 15.23 = 33.86 kN/m.
Since the natural frequency of a spring system varies as the
square root of the effective spring rate, the natural frequency
for single wheel bump at the front is 35% higher than the
frequency for double wheel bump. The rear anti-roll bar is
not very stiff and the single wheel bump frequency at the rear
is only about 14% higher than the double wheel bump
frequency.
210 A high-performance sports car: Porsche 928
The noticeable hardening of the front suspension under
single wheel bump has received some adverse comments
from a few members of the motoring press. On the whole, it
appears to be a small price to pay for a car that rides so well
and comers like a real sports car - and that undoubtedly is
the best description of the Porsche 928.
Index